There's some sampling bias at play because you don't hear about the less flashy examples. I use machine learning for particle physics, but there's no marketing nor outrage about it.